ORDINALS FEAR & GREED INDEX BETA

METHODOLOGY - FULL TRANSPARENCY, NO BLACK BOX

Update, Nov 1 2026: after 30 days of data the index was moving mostly on noise: its daily changes alternated up and down, and the label switched between Fear and Extreme Fear 12 times in 22 days. Volume, sales, listings and bids now use 7-day figures instead of a single day, and Floor Price Trend averages the middle half of the 32 collections instead of taking the median. Weights are unchanged. Every day from Oct 9 was recalculated; the previous values are kept in the public JSON (indexV2). On the same data the typical day-to-day move went from about 7 points to under 2.

Update and correction, Oct 22 2026: ord.net lists inscription-number groups (sub100, sub1k, sub10k, sub100k) next to real collections, and they overlap: every sub100 sale is also a sub1k, sub10k and sub100k sale. Adding all four up counted one trade up to four times. On Oct 20 a single 0.23 BTC sale was counted as 0.92 BTC, which pushed that day to 34. The three inner groups are now left out, and Oct 20 was recalculated to 27 (the original value is kept as indexBeforeCorrection in the public JSON). Also new: Shielded Ordinals data (Ord Dropz's private-ownership protocol, whose trades don't show up on ord.net or satflow.com) is now collected in the background every day. It is not part of the score yet; it will be judged once it has a few weeks of history.

Update, Oct 16 2026: Floor Price Trend, the new Bid Support component and Listing / Delisting now read both ord.net and satflow.com. Marketplaces often show slightly different floors and bids, so the two are never averaged: each collection uses the lowest floor and the highest bid found on either one (the price a buyer or seller really gets), and each 7-day or day-over-day change compares a marketplace only with itself. Floor Price Trend went from 20% to 15%, Bid Support takes the 5%. Earlier days are not recalculated.

Methodology updated on Oct 9 2026; history recalculated. Mint Velocity and New Supply Growth were removed from the score (they are still shown below, for information only). Floor Price Trend and Volume Momentum were rebuilt. Days before Oct 9 keep the floor and volume scores they originally had, because the data needed to recompute them the new way was not stored; the recalculated history and the original values (indexLegacy) are both in the public JSON.

The index is a weighted average of six scored components, each measuring a different facet of Ordinals market behavior. Two more components are shown for information only and carry no weight. Components are scored in one of three ways. Volume Momentum and Sales Velocity compare their 7-day average with their own 30-day average, and Floor Price Trend and Listing / Delisting Ratio map a 7-day change straight to a 0–100 score; none of them needs any warm-up. Bid Support is percentile-ranked against this index's own trailing history, so its score reflects "unusually high/low compared to its own recent past", not a fixed threshold. Market Cycle Position ranks against a multi-year baseline going back to February 2023 from day one - see below.

Final index = Σ(score × weight × confidence) ÷ Σ(weight × confidence), rounded to the nearest integer. 0–24 Extreme Fear, 25–44 Fear, 45–55 Neutral, 56–75 Greed, 76–100 Extreme Greed.

Confidence weighting: a percentile-ranked component with only 2-3 days of trailing history is statistically unreliable - with that few data points, percentile rank has only 2-3 possible outcomes, so a trivial move can land at 0 or 100 and swing the whole index on noise. Each of those components' weights is scaled by min(days of history ÷ 14, 1), then all weights are renormalized back to 100% - so a thin history shrinks a component's real influence instead of quietly getting full voting power it hasn't earned yet. In the Components list below, a weight shown as "12% → 7%" means the component is still warming up. Volume Momentum, Sales Velocity, Floor Price Trend, Listing / Delisting and Market Cycle Position start at full confidence immediately: the first four are scored on fixed scales against this index's own 7- and 30-day figures, the last against its 43-month baseline.